Development of wearable multi-lead ECG measurement device using cubic flocked electrode

Development of wearable multi-lead ECG measurement device using cubic flocked electrode

Cubic flocked electrode

Figure 1a shows the CFE fabrication process using electrostatic flocking technology, where Ag-plated fibers were formed on the surface of a urethane form. The fibers were charged at 50 kV. Therefore, the fibers stand vertically against the urethane form surface, as shown in Fig. 1b. The CFE has dimensions of 20 mm × 20 mm and is 12 mm thick. The hardness of the CFE is 20° (SRIS 0101/Asker C). The flocking process is described in detail in the “Methods” section.

Figure 1
figure 1

Picture and fundamental evaluation of the cubic flocked electrode fabricated in this work. (a) Image of CFE fabrication process using electrostatic flocking technology. (b) Photograph of the CFE and cross-sectional view (at the A-A’ line). (c) Resistance changes in the CFE when compressed in the directions of the x, y, and z axes. (d) Impedance spectra of the CFE, PEDOT:PSS electrode, and Ag/AgCl gel electrode, measured in the 1 Hz to 10 kHz frequency range when in contact with skin (forearm) at 200 Pa. The spectra show the mean values and standard errors (measurements were taken five times).

First, we evaluated the fundamental characteristics of the CFE. Because the CFE is flexible, it can be deformed greatly by compression. The effect of this deformation on the resistance value of the CFE itself was therefore evaluated. Figure 1c shows the changes in the resistance value when the CFE is compressed in the three axial (x, y, and z) directions. The compression rate ranged from 10 to 50%. The resistance value fluctuation is the largest along the z-axis (from 14.1 Ω at 100% to 8.0 Ω at 50%). The resistance fluctuations along the x and y axes were smaller than those along the z-axis (less than 10 Ω). However, when compared with the skin-electrode impedance, which is approximately a few megaohms, the resistance change was small enough to be negligible. Therefore, large-scale deformation of the CFE itself does not affect the ECG measurements.

Figure 1d shows the skin-electrode impedance characteristics of the CFE, a poly(3,4-ethylene dioxythiophene) polystyrene sulfonate (PEDOT: PSS) electrode (C3fit IN-pulse, Goldwin Inc., Japan), and an Ag/AgCl electrode (Red Dot, 3M, USA) with gel. The measurement frequency range for the impedance is from 1 Hz to 10 kHz and the measurement position was at the forearm, as shown in the figure29. Because the impedance is dependent on the contact pressure28, the contact pressure between the skin and the electrode was fixed at 200 Pa. From this result, the impedance of the CFE was higher than that of the Ag/AgCl electrode with gel. However, when compared with the PEDOT: PSS electrode, which is an industrial product, the impedance was at the same level for both electrodes. This result shows that the impedance is small enough to allow the CFE to be applied to ECG measurements.

MA reproduction experiment

To evaluate the MA durability of the CFE, an MA reproduction experiment was conducted using a skin phantom. The purpose of this experiment is also to obtain the appropriate contact pressure between the skin and the CFE for the design of the wearable multi-lead ECG measurement design. The magnitude of the motion artifact Vma is given by the following equation30.

$${\mathrm{V}}_{\mathrm{ma}}=2\left[\Delta {\mathrm{V}}_{\mathrm{dc}-\mathrm{offset}}+\Delta {\mathrm{Z}}_{\mathrm{skin}-\mathrm{electrode}} \left(\frac{\left(\frac{{\mathrm{V}}_{\mathrm{ECG}}}{2}\right)+{\mathrm{V}}_{\mathrm{dc}-\mathrm{offset}}}{{\mathrm{Z}}_{\mathrm{in}}}+{\mathrm{I}}_{\mathrm{bias}}\right)\right]$$


Vdc-offset, Zskin-electrode, Vecg, and Ibias are the offset voltage generated at the skin-electrode interface, the impedance between the skin and the electrode, the ECG signal voltage, and the input bias current of the amplifier, respectively. ΔVdc-offset and ΔZskin-electrode are major causes of MAs. The purpose of this paper is to reduce both ΔVdc-offset and ΔZskin-electrode, and to show experimentally the type of pressure and conditions under which Vma can be reduced by using a raised electrode.

Figure 2a,b show the experimental system. The skin potential of the right arm (RA) position, which was generated using an ECG simulator (ProSim8, Fluke Corp., USA), was connected to the skin phantom (HXBNXTB858510MX, Wet Lab Inc., Japan). The CFE was attached to the surface of the skin phantom and was then connected to an ECG amplifier (Analog Devices, USA). The CFE was glued to a textile (MCM3749, Under Armour, USA), which is the same material that is used for the wearable multi-lead ECG measurement device described later in the paper. The edge of the textile was fixed to a jig. The contact pressure between the electrode and the skin phantom was measured using a force gauge. Additionally, the skin potential of the left leg (LL) was connected to the amplifier directly. The ECG signal was recorded using an oscilloscope as the output ECG signal. The skin phantom was attached to an x, z, and θ axial automatic stage. This automatic stage moved in a round trip such that displacement occurred between the electrodes and the skin phantom. As a result, an MA was reproduced. Note that the center of the axis of rotation of θ is the contact point between the skin phantom and the electrode.

Figure 2
figure 2

MA evaluation when measuring ECG using a skin phantom. (a) Experimental setup of MA evaluation equipment, and (b) photograph of contact part. (c) Experimental results of the MA evaluation experiments when the automatic stage displaced the x-axis (Δx = 500 µm, 1000 µm, 1500 µm), the z-axis (Δz = 500 µm, 1000 µm, 1500 µm), and the θ-axis (Δθ = 0.5°, 1.0°, 1.5°). Sample A is the textile on which the flocked electrode was formed directly. Sample B is the CFE, which was attached to the textile.

Two samples were prepared to compare their ECG signal quality levels. Sample A is a flocked electrode that was formed on the textile directly. Sample B is the CFE attached to the textile, as shown in Fig. 2c. These samples were placed in contact with the skin phantom such that the fixed part of the cloth was aligned with the surface of the skin phantom. This contact state is a reproduction of the contact state of the actual wearable device itself. The contact pressures measured using the flocked electrode sample and the CFE sample were 57.3 Pa and 495.2 Pa, respectively, as calculated using the output of a force gauge. The ECG signal was measured when the automatic stage reciprocated on the x-axis (Δx = 500 µm, 1000 µm, 1500 µm), the z-axis (Δz = 500 µm, 1000 µm, 1500 µm), and the θ-axis (Δθ = 0.5°, 1.0°, 1.5°). These displacement values were calculated based on the chest expansion rate during respiration31. The frequency of the stage motion was 14 round trips/min, which was equivalent to the average human breathing frequency. The measurement time was 20 s.

About the results for Sample A, the largest MA was observed when the stage moved along the x-axis. The MA is confirmed, even at Δx = 500 μm. This is because of the occurrence of slippage. Because the initial contact pressure was 57.3 Pa, the flocked electrode could not follow the x-axis displacement, and the skin phantom-electrode interface slipped as a result. For the z-axis and the θ-axis, the MA was suppressed because the textile absorbed the displacement, which meant that the MA was smaller than the x-axis displacement. In contrast, in the Sample B results, MAs were hardly observed on the x-axis, the z-axis, and the θ-axis. It is believed that this result was caused by the increased contact pressure when making the cubic shape of the urethane form. This means that the slippage was suppressed by the shear deformation of the urethane form. The results above show that the occurrence of MAs can be reduced by using the CFE. Furthermore, we obtained a design indicator that contact pressure of 500 Pa or more can reduce the number of MAs being caused by respiration.

Wearable multi-lead ECG measurement device

Figure 3a shows the wearable multi-lead ECG measurement device that we developed in this work. The view is from the rear side of the device. Compression wear (MCM3749, Under Armour) was used as a substrate textile to increase the contact pressure between the skin and the CFE. On the compression wear surface, 18 CFEs were attached around the chest area, and four CFEs were attached to the limbs using adhesive paste. The positions of these electrodes were designed to measure the V1–V6 signals of the Mason-Likar (ML) lead ECG signals, which are equivalent to the V1–V6 signals of the 12-lead ECG signals32. The 18 electrodes located around the chest can be divided into six upper electrodes, six middle electrodes, and six lower electrodes. The reason for this division procedure is to reduce the influence of electrode positioning errors due to differences in the way that individual patients wear the device. These electrodes were connected to the pads in the stomach position using conductive thread (AGposs, Mitsufuji Corp., Japan) and were sewn in position using a sewing machine. The pads were also fabricated using the electrostatic flocking technology and were connected to corresponding pads on the front side by through-wiring. The conductive threads were protected and covered with a film to prevent the threads from attaching to the patient’s skin directly because direct attachment can cause noise on the ECG signals. The measurement system is described in detail in the “Methods” section. Figure 3b shows the evaluation of the contact pressure between the CFE surface and the skin of a torso (CPLM, TOMANE CO., LTD., Japan). The contact pressure was measured by the insertion of a balloon that was connected to an air pressure measurement system (AMI3037-2, AMI Tec Inc., JAPAN). To optimize the contact pressure according to the MA evaluation experiment, the CFE’s height was designed by checking several contact pressures in order up to more than 500 Pa. Finally, the heights of the CFEs were determined as shown in Fig. 3b. The average contact pressures with the optimized CFE at positions V1–V6 were 0.6 kPa, 0.53 kPa, 0.53 kPa, 1.83 kPa, 0.37 kPa, and 0.73 kPa, respectively. These results show that the contact pressures are almost all more than 500 Pa, which was indicated in the MA evaluation results.

Figure 3
figure 3

Multi-lead ECG wearable measurement device. Photographs of (a) rear side, and (b) front side of the wearable device. (c) Contact pressure distribution on the CFE, which was attached between the wearable device and the torso.

Figure 4a shows a photograph of the ECG measurement experiment with the wearable multi-lead ECG measurement device. The experimental protocol was as follows.

  1. 1)

    Put on the wearable device without gel.

  2. 2)

    Measure the multi-lead ECG signal in the spine position while breathing freely (2 min).

  3. 3)

    Take off the wearable device.

  4. 4)

    Measure the 12-lead ECG signal using a medical product (Smart ECG, ECG Lab Corp., Japan) in the spine position while breathing freely (2 min). The electrodes were placed by a medical doctor. The number of subjects for this experiment was 10.

Figure 4
figure 4

Experimental results of multi-lead ECG measurements. (a) Photograph of the experiment. The subject put on the wearable device and rested in the supine position while breathing freely. (b) ECG signal (from V3 position of middle electrodes) from a 2 min measurement period.

Figure 4b shows the ECG signals acquired in 2 min at the V3 position of the middle six electrodes. Both signals with MAs and signals without MAs can be observed from the ECG signals of this 2 min period. In a signal without MAs, the P, Q, R, S, and T waveforms can be observed clearly. In contrast, the signal with the largest MA in this experiment was detected as shown in the upper right area of this figure. From this signal, while the R waveform is the exception, the P, Q, S, and T waveforms could not be observed because of the fluctuations in the baseline. However, these fluctuations occurred only six times in 2 min (marked by red triangles). Furthermore, the fluctuations occurred for only a few seconds and soon disappeared. This result indicates that the cause of the baseline fluctuation was a slight movement of the human body rather than breathing.

Figure 5 shows the ECG signals measured by the wearable multi-lead ECG measurement device (18 electrodes) and by medical products (six electrodes). Figure 5a shows the ECG waveform of a subject who succeeded in producing a stable measurement. About the 18 lead ECG signals, we successfully acquired waveforms of sufficiently good quality to allow the P, Q, R, S, and T waves to be identified in all leads. In addition, the ECG signal waveforms for each lead are similar to those of the six-lead ECG signal obtained using the medical product. This confirms that the proposed ECG offers the same quality as the medical product. In contrast, Fig. 5b shows the ECG waveform for a subject whose measurements were not stable. Fluctuations of the baseline and noise are observed in signals V1, V2, and V6. Signal V1 of the upper electrode cannot be confirmed, even with the R waveform. In contrast, in signals V3 to V5, we have achieved stable ECG signal measurements to some extent. Because the subject was slender (chest circumference: 86.0 cm) when compared with the other subjects (average chest circumference: 92.2 cm), it is considered that stable ECG measurements are possible, even in the V1, V2, and V6 signal areas, by designing wearable multi-lead ECG devices to be a specific size. The results of the demonstration experiment above show that it is possible to measure an ECG at the same level as a medical product with the patient in a resting state by using the proposed wearable multi-lead ECG measurement device with the CFE.

Figure 5
figure 5

The 18-lead ECG signals measured by the wearable multi-lead ECG measurement device and by a commercial medical product around the chest. (a) Best quality ECG from 10 subjects, and (b) worst quality ECG from 10 subjects.

#Development #wearable #multilead #ECG #measurement #device #cubic #flocked #electrode

Related Articles

Leave a Reply

Your email address will not be published. Required fields are marked *

Back to top button